DATA AND NOTES. 351 



Viti: ika, fish. Rotuma: i'a, id. 



Epi: yika, fish. Sesake, Ambrym, Santo, Southeast Epi: ika, id. 

 Bierian: ne-ika, id. Mota, Lo, Fagani, Belaga, Nggela: iga, id. 

 Arag, Omba: ige, id. Wango, Bululaha: i'a, id. Ugi, Ulawa, 

 Alite: ia, id. Roro: maia, id. Norbarbar: ie, id. Nengone: 

 wa ie, id. Saa, Tubetube: He, id. New Georgia: ihana, id. 

 Dobu, Tavara, Awalama, Taupota, Wedau, Boniki, Mukawa, 

 Kubiri: iana, id. Kabadi: veana, id. Duke of York: ian, id. 

 Kiviri : iun, id. Kiriwina : iena, id. Murua : iini, id . Matupit : 

 ten, id. Baravon: en, id. Buka: eina, aenna, aienne, aiena, id. 

 Merlav, Pak: ig, id. Motlav, Gog, Mosin, Alo Teqel, Volow: 

 eg, id. Malekula: na-ih, id. New Ireland (Duffield) : kauk, id. 



North Borneo: jikan, fish. Malay, Massaratty, Teor, Ilocan, 



Mayapo, Gah : ikan, id . Bouton : ikani, id . Amblaw : ikiani, id. 

 Menado: maranigan, id. Silong: ackan, id. Sulu: m?a, id. 

 Kar Nicobar: &a, id. Central Nicobar: ga, id. Lariko, 

 Wahai, Gani, Saparua, Ahtiago, Matabello: ian, id. Liang, 

 Morella: iyan, id. Tidore: nyan, id. Batumerah, Awaiya, 

 Caimarian: iani, id. Mysot: «'n, id. Java: iwa, id. Teluti: 

 yano, id. 



Hebrew: da#, d&fo dagah, degath, fish. 



In the data here assembled we have two working stems, which, from the 

 regions of their greatest frequency, are to be described as ika the Polynesian 

 stem, and ikan the Indonesian stem. 



We find ika as the stem in all but five of the Polynesian languages ; and 

 the only change affects the consonant; in Samoa and Rotuma it has so 

 lately vanished that enunciation still marks the gap ; in Nukuoro, Tahiti, 

 and Hawaii it has been forgotten that a consonant ever intervened. 



This stem is of the most common occurrence in Melanesia. In various 

 stages of dilapidation we may trace it from Nengone in the far south to 

 the central Solomons and with possibly one sporadic instance farther north 

 in the eastern gateway. Epi shows in yika the ika stem with a slight semi- 

 vowel reinforcement, possibly repeated in North Borneo jikan, though the 

 value of this j is not distinctly set forth. The normal ika is found in the 

 New Hebrides, and with a variant not possible in Polynesia, as iga and 

 ige in the New Hebrides and the central Solomons. The change of the 

 final vowel from a to e is so slight as not to call for detailed consideration. 

 Wango and Bululaha reproduce Samoa and Rotuma i'a. Ia is found in a 

 part of the Solomons, but not restrictively, for ie carries it along to Norbar- 

 bar and Nengone. We are last to consider a variant which would be im- 

 practicable in Polynesia, terminal abrasion to a closed stem, ig and eg in the 

 Banks Group, ih in Malekula, perhaps in the New Ireland composite ka-uk. 



Modifications of the ikan stem are found in the northern part of Mela- 

 nesia, the southernmost occurrence as well as the least altered being New 

 Georgia ihana. The presence of the n serves to identify ikan for us, despite 

 the dropped k, in Duke of York, Matupit, Buka and Baravon. These are all 

 within the range of Post- Polynesian Malayan voyaging, recourse to which 

 explanation has been had in jolau (287). 



